Search

As is known, cable plays a significant role in electric power transmission of the ship. From the safety perspective, this paper aims to predict the ship cable life more accurately and analyze its reliability under small sample conditions. To achieve this goal, based on the non-equidistant grey model (NEGM (1, 1)), an optimized equal-dimension dynamic NEGM (1, 1) is proposed in this paper. In the improved model, the raw data smoothness and the background value are optimized to apply to grey model parameters. Equal-dimension dynamic prediction algorithm (EDDPA) is adopted to predict the lifetime of ship CXF type cable. Compared with the aging model from literature and NEGM (1, 1), the prediction effectiveness and reliability of the proposed model are more considerable. Interestingly, with the increase of recurrence times, the numerical results show this model has higher prediction performance. Finally, the reliability and confidence interval of the life evaluation of ship CXF cable are given to provide more valuable information for the cable management. [ABSTRACT FROM AUTHOR]

This paper presents the three-phase impulse overvoltage measured at the terminal of a 10 kV distribution line when the impulse current was injected into the phase conductor of the distribution line. The overvoltage observation station was installed at the terminal of the line to record the three-phase overvoltage waveforms. It is shown that the struck phase overvoltage consists of three components including the negative pulse, the stage of residual voltage maintenance and the overshoot oscillation part. The observed negative peak values of the struck phase overvoltage range from −85.4 kV to −96.0 kV in three different injected impulse levels at 90,120,150 kV charging voltage. The duration of residual voltage depends on the duration time of injected current. Furthermore, there is a strong positive relationship between the oscillation amplitude of the struck phase overvoltage and the injected current amplitude. The peak amplitudes of three-phase overvoltage always have the following magnitude relationship: struck phase C > induced phase B > induced phase A. The voltage coupling coefficients of induced phases will not exceed 0.4. The test method proposed in this paper will be helpful to the research on the lightning overvoltage response characteristics of the distribution line and the protection analysis of arresters. [ABSTRACT FROM AUTHOR]

This paper reports the waveform characteristics and parameters of the overvoltage at the terminal of a 1.5 km 10 kV distribution line by direct triggered lightning. The three-phase overvoltage exhibited three kinds of waveforms, all including the initial rising process of the negative peak, the transition from negative peak to the positive peak or the residual voltage, and the subsequent oscillation process. It is observed that the negative peak value of the three-phase overvoltage is between −8.2 kV and −96.1 kV and the rise time of negative peak is between 3.2 μs and 31.7 μs, while the corresponding peak value of the return stroke current is between 3.3 kA and 32.9 kA. When the return stroke current is relatively small, the negative peak value of the direct struck phase overvoltage is twice or more than that of the indirect struck phase. When it is larger, the negative peak value of the indirect struck phase overvoltage is much closer to that of direct struck phase. There is good linear relation between the negative peak value of the indirect struck phase overvoltage and the peak value of the lightning return stroke current, with correlation coefficient being 0.98 and 0.90. [ABSTRACT FROM AUTHOR]

Appropriate shielding methods can reduce the lightning induced voltage and current of cables. In this article, different lightning surge currents (8/20 and 5/320 μs) are coupled to the tested multi-core cable with single shielding layer and double shielding layers. The shielding effectiveness (SE) on the lightning surge current is compared for different cable shielding methods. The experimental results show that the cable induced current is effectively reduced when the shielding layer, the shielding spare core wire or the metal tube are grounded at two ends. A better shielding effect for the cable with double shielding layers is achieved by grounding both ends of the outer shielding layer and one end of the inner shielding layer, reducing the induced current to about 5% when the shielding layer is not grounded. Moreover, the SE of the spare core wire, used as the shielding wire in the multi-core cable, is better than that of the external shielding wire of the cable. The SE of the metal tube increases more and more slowly with the increase of the tube coverage percentage. Furthermore, the descending order of the SE of shielding methods involved here are grounding both ends of double shielding layers, grounding both ends of single shielding layer, spare core wire shielding and metal tube shielding. Finally, some engineering recommendations are proposed through the comparison between the shielding methods in this article and relevant works. [ABSTRACT FROM AUTHOR]

The optical return-stroke speed calculation is vital to analyze the lightning development process. The traditional calculation methods of lightning optical return-stroke speed, including the slope-intercept and percentage-of-peak methods, usually have considerable calculation errors in noise cases. In this article, we generate the simulated optical return-stroke signal within 100 m above the starting point along the lightning channel and use four cross-correlation methods and six traditional methods to calculate the return-stroke speed at different observation points along the channel. We optimize the cross-correlation method and reduce the calculation error to within ±10% by selecting a part of the waveform to participate in the cross-correlation operation and performing the exponential operation on the frequency-domain cross-correlation sequence. The calculation results of the quadratic-correlation method and two optimized cross-correlation methods do not change significantly at different observation points along the channel. Different channel-base waveforms affect the return-stroke speed variation trend. The smaller the dispersion factor λp is, the more pronounced the variation of the calculation result is, and the greater the calculation error is. The optical waveform with generated Gaussian noise will significantly affect the traditional calculation method. In contrast, the cross-correlation calculation method with noise has only a tiny calculation error. [ABSTRACT FROM AUTHOR]

This article presents the electric field waveforms of natural return strokes recorded by Foshan Total Lightning Location System. The distance ranges from 10 to 300 km. The evolution of parameters for positive first and negative first and subsequent return stroke electric field waveforms along with distance change is analyzed. The results show that the risetime, fall time, half-peak width, and zero-crossing time become longer with longer distance, and these four parameters of the negative return strokes show more clear linear relation with distance than the positive return strokes. For the overshoot duration, the values for both positive and negative return strokes increase with distance at 10–200 km, but when the distance is larger, the distance dependence becomes less obvious. The values of the initial peak normalized to 100 km and 20 kA remain the same in general for all the return strokes. For the overshoot amplitude and the ratio of overshoot to initial peak, the values for both positive and negative return strokes increase with distance. The return stroke electric field waveforms at far distances are more likely to show clear opposite polarity overshoot. [ABSTRACT FROM AUTHOR]

This article presents an experiment of rocket-triggered lightning to 10 kV power distribution lines and the characteristics of lightning current at Guangzhou Field Experiment Site for Lightning Research and Testing during 2018–2019. The return strokes (RSs) current of the triggered-lightning flashes in this study exhibits a shorter half-peak width and similar risetime and peak current compared with those of subsequent strokes current from natural lightning. The peak current of RSs shows strong power function correlations with charge transfer within 1 ms and action integral within 1 ms of RSs, with R2 of 0.80 and 0.92, respectively. Besides, the peak current of RSs is proportional to the average rate rise of RSs (dI/dt). We recommend the double HEIDLER function to simulate the RS current and provide accurate raw data for the study of lightning protection. [ABSTRACT FROM AUTHOR]

The impulse grounding performance of distribution lines tower is the key factor to lightning protection. Few field high-current impulse test researches are conducted on distribution line tower. In this paper, we used a mobile high-current impulse generator to conduct test on a 10 kV distribution line tower. The peak value of the current injection is 6 kA to 16 kA. The impulse impedance and the surface potential distribution are measured and analyzed. Based on the CDEGS software, an equivalent modeling of the experiment was carried out. The influence of soil ionization was taken into account by the method of changing the conductor radius through two rounds of calculation. The simulation and measured results were basically consistent. Experiments have shown that the soil ionization effect has an important impact on lightning performance of distribution line tower. The graphite earthing modules (GEMs) and pulling wires can significantly reduce the impulse impedance of the tower. The potential rise of the earth surface at a distance of 30 m from the tower can reach 1 kV at peak, under an injection of 7.15 kA. [ABSTRACT FROM AUTHOR]

In this article, we analyzed the characteristic parameters of close vertical electric field waveforms from triggered lightning at 58 and 90 m obtained at Conghua, Guangzhou, in the summer of 2019. A total of 11 characteristic parameters were defined, including leader electric field peak [kV/m] (EL), return stroke electric field change [kV/m] (ERS), zero-to-peak rise time [μs] (T1), 10-to-90% rise time [μs] (T2), half-peak width [μs] (T3), zero-crossing time [μs] (T4), slow-front duration [μs] (T5), opposite polarity overshoot duration [μs] (T6), fast-transition 10-to-90% rise time [μs] (T7), opposite polarity overshoot ratio [%] (EOS/EL), and slow-front magnitude relative to peak [%] (ES/EL). The arithmetic mean, geometric mean, and standard deviation of each parameter were counted. Data comparison results at 58 and 90 m show that T1–T7 at 90 m were larger than those at 58 m. EL and ERS at 90 m were both positively correlated with the peak current [kA] (IP). The influence of distance on this correlation was also analyzed. The half-peak width of the current waveform was inversely proportional to that of the electric field waveform. [ABSTRACT FROM AUTHOR]

In this article, we analyzed the difference in the estimated currents from the Foshan Total Lightning Location System (FTLLS) under two triggered lightning situations, i.e., triggered lightning to ground and triggered lightning to the distribution line. FTLLS underestimated the peak currents of triggered lightning to the distribution line, which is reflected in the peak far electric field. The peak far electric field of triggered lightning to the distribution line is about 12% lower than that of triggered lightning to ground. However, this difference in estimated current is not reflected in the far electric field waveform parameters, at least for the three field parameters: zero to peak risetime, 10%–90% risetime, and half-peak width time. The presence of distribution line will increase the risetime of the lightning return stroke current, the arithmetic mean of the current risetime of triggered lightning to distribution line is 0.46 μs, and that of triggered lightning to ground is 0.27 μs. Clarification of this underestimation phenomenon is beneficial for the application of the lightning location system in power systems. [ABSTRACT FROM AUTHOR]

This article presents the electric field characteristics of 1612 single, 912 first, and 2901 subsequent return strokes (RS) in negative lightning in Guangdong Province, China. Statistic parameters of electric field waveforms include risetime, fall time, half-peak width, zero-crossing time, overshoot duration, initial peak, overshoot amplitude, and ratio of initial peak to overshoot (Ep/Eos). The results show that the electric field parameters of single RS are similar to those of the first RS. The difference of the geometric mean (GM) for all the parameters is within 10%, except for the Ep/Eos. The value of Ep/Eos for the first strokes is 14.3% larger than that of the single RS. But the values of these parameters for subsequent RS show significant differences from the single and first RS. The values of overshoot duration for subsequent RS are similar to those of the first RS, while the ratio of initial peak to overshoot is larger. As for the other parameters, the arithmetic mean and GM values are all more than 40% larger than the corresponding values of subsequent RS. [ABSTRACT FROM AUTHOR]

The effects of long upward connecting leader (UCL) channel on electromagnetic fields for lightning striking to tall towers are studied via lightning transmission-line return-stroke model. The current distributions along the lightning channel and the tower are derived, and the resultant lightning electromagnetic fields with various UCL lengths above tower are calculated at close and far distances. It can be found that the contribution of the various components of the electric and magnetic fields depends strongly on the distance to the observation point. Electric field has a shading effect showing that the field amplitude generated by lightning striking to tall towers is smaller than that to ground, and that this shading effect depends on the UCL length and the tower height. However, we cannot find such an effect for the magnetic field. The shading effect of the electric field increases with the increase of the junction height above tower, but disappears when the observation distance exceeds a critical distance. Modeled results are also compared with their measured counterparts associated with lightning striking to the 600 m tall Canton Tower. The proposed model can explain why the ratios of peaks of luminosity to magnetic field of the downward first stroke are considerably larger than those of the subsequent strokes as being due to the fact that the junction height above tower for first stroke is higher than that for subsequent stroke, whereas the downward return stroke speed of first stroke is smaller than that of subsequent stroke. [ABSTRACT FROM AUTHOR]

Insulators are one of the widely used important devices in power systems, and proposing an accurate and effective online monitoring method for insulator defects is an urgent need for intelligent and informative power systems. We propose a multiscale residual neural network (MSResNN) for insulator surface breakage recognition. The proposed multiscale convolution filter uses three different size convolution kernels for convolution filtering and performs feature maps fusion, which can enrich the spatial correlation and channel correlation of feature maps. We investigated the effects of different learning rates on the model performance and confirmed that the model performs best with learning rate of $10^{-6}$. Compared with VGG16, Inception_v3, and Resnet50, the MSResNN model has significant advantages in detection capability and occupies less computational resources, and the detection speed can meet the industrial real-time detection requirements. Through dark–light and noise experiments, it is demonstrated that the proposed method is robust and can effectively deal with low-quality insulator images. [ABSTRACT FROM AUTHOR]

As a kind of transformer installed on vehicles, the vehicular mobile transformer needs to realize frequent transportation quickly and safely. However, there are few studies on the on-line monitoring of vehicular mobile transformers. In this article, a transportation condition monitoring system for vehicular mobile transformers is introduced and developed. The transportation process of a 110-kV/40-MVA vehicular mobile transformer is monitored from Jiangsu Province to Guangdong Province, which is recognized as the first time that the transformer is monitored for more than 12 h and more than 1000-km transportation. The geo-location, driving speed, vibration, and shock acceleration in six different road sections are obtained in real-time. The results show that the vibration and shock acceleration has a strong positive correlation with the driving speed. The amplitude of the $y$ -axis acceleration is generally the lowest among three-axis accelerations, and the three-axis vibration and shock acceleration has the same trend of increase and decrease. During the whole transportation process, the driving speed of the 110-kV/40-MVA vehicular mobile transformer varies from 0 to 76 km/h, and the vibration and shock acceleration amplitudes obtained by the monitoring system do not exceed 1 g, which has great significance to the safe operation of power grid and can provide the strong evidence of online monitoring data to set transportation guidelines for vehicular mobile transformers. [ABSTRACT FROM AUTHOR]